Analysis Tool Generates Custom Vehicle Drive Cycles Based on Real-World Data
September 26, 2011
Understanding vehicle usage plays a fundamental role in assessing the performance of new vehicle technologies, leading to more informed decision making, better test procedures, more successful designs, and lower manufacturing and operating expenses. The U.S. Department of Energy’s National Renewable Energy Laboratory (NREL) recently launched a new analysis tool that produces representative, testable drive cycles at record speed from large amounts of vehicle data gathered using onboard logging devices.
The new Drive-Cycle Rapid Investigation, Visualization, and Evaluation (DRIVE) tool uses global positioning system and controller area network data to produce custom vehicle test drive cycles based on real-world activity, analyzing thousands of hours of data in a matter of minutes. The tool also provides a wealth of useful statistics on the data set and searches for an existing industry-accepted drive cycle that provides the best fit for the supplied data. Algorithms mimic the logic and expertise of a human engineer, cutting testing and analysis time by days or weeks, while establishing a repeatable process and making information accessible through a simple graphical user interface.
Originally geared toward analysis of medium- and heavy-duty vehicle fleets, use of the DRIVE tool has expanded to encompass the full range of vehicle types and sizes.
“We’ve received great feedback from fleet managers and vehicle developers alike,” said NREL's Adam Duran. “Fleet managers are excited about using it to make educated decisions about how to best invest in hybrid technologies, for instance, because they can determine in advance where the technology might have the optimal payback period. Vehicle developers can use it for more efficient and accurate vehicle component sizing, reducing technology cost and development time.”
The DRIVE tool helps manufacturers ensure designs are based on real-world usage, supplying information needed to perform vital development tasks, such as sizing electric motors in a hybrid vehicle configuration or optimizing battery storage in an electric vehicle. For researchers this information can lead to improved models and more precise experiment designs. For regulators, a better understanding of the way people drive can help guide policy regarding vehicle emissions, incentives, and fuel economy test procedures.
NREL’s tool filters large sets of raw data, removing erroneous data points and repairing missing data sections, before performing analyses covering 168 unique vehicle drive cycle metrics. The program then generates shortened custom drive cycles from “ideal” sections of filtered data using specialized statistical clustering methods. In addition, the program compares filtered in-use data to a library of standard test cycles to find the best fit. Output results range from simple tabulated summary statistics to Google Earth route maps, providing information that has enough depth for scientific applications, but is accessible to users without technical backgrounds.
“While we developed DRIVE for the Department of Energy to better understand and evaluate vehicle drive cycles, its methods and algorithms could potentially be adapted for a wide range of alternative applications,” Duran said. “Understanding real-world cyclical thermal stresses, solar availability, battery charge/discharge cycles, and bending stresses are just some of the other potential platforms of interest that could be explored.”
“The key to DRIVE is capturing the real-world data of interest, determining important metrics, and then applying the unique methodology and algorithms to generate custom test cycles for accelerated testing and simulation,” he added.
For more information about the DRIVE tool, contact NREL's Adam Duran.